The State conducts administrative case reviews on each child in physical or legal custody of DCF, including children in their own homes without court jurisdiction. According to State policy, the initial administrative case review ("ACR") is conducted within 45 days of the initial placement and every 6 months thereafter.

LINK currently does not have the capability to collect and report the date of the most recent periodic review. The current AFCARS extraction program derives this date from the court disposition date. If there is no court disposition date, the program code subtracts one month from the end of the report period date and enters this as the review date.

Case File Review Finding: See the findings in section 3 of Tab A. There were 39 records that did not match what was in the AFCARS file. The most significant finding was that there were 14 records that had dates of August 30, 2000 in the AFCARS file, and of those the reviewers found dates of review that occurred prior to that date in seven records. In another 2 records, there were review dates after the August date. There were seven records marked as questionable because they had 8/30/2000 as the review date in AFCARS.

There were 14 records that had dates other then 8/30/2000, but reviewers did not find these to be the correct review dates.

There were three records that the reviewers noted that a case review was not required because the child was over the age of 18.

Post-site visit Information: During a post-site visit follow-up, the State confirmed that the August 2001 implementation of a new treatment plan module did not include LINK programming to collect the periodic review date. The State shared that the periodic review date will be collected with the implementation of the "DCF-553", which is now on hold.

This is a mandatory field. In order for caseworker to create a removal episode, the date of birth for the child must be entered into the system.

During intake, when the child's date of birth (DOB) cannot be established, e.g. an abandoned child, the informal policy is for workers to enter January 1 of the estimated year of birth. Typically, treatment workers update a child's estimated DOB when more information on the child becomes known.

Per AFCARS guidance, the State should instruct workers that, in cases where the child's date of birth is unknown, use the 15th day of the estimated month.

LINK data entry screens and program code allows the collection of multiple races. This is a mandatory field that must be completed by the investigation caseworkers. The investigative case cannot be closed without information on child's race being entered.

The frequencies for race indicate a large percent (24%) of children with a race of "unable to determine." The "unable to determine" response is appropriate if a child was abandoned or if the child, parent and/or caretaker refuses to specify race. The State team agreed that it may also be possible that workers are entering "unable to determine" as the child's race when the child is identified as being of Hispanic or Latino origin.

Also, the AFCARS Foster Care Design Document states that if the "cd_race" field on the person table is blank, then substitute the value to indicate "white". The State needs to check to make sure the program logic is not mapping missing data to "white". The Design Document is not the current version, as noted by the use of "1" as the AFCARS value for "white". The document "Additions to Detailed Design" does not include this reference.

On the screen that contains the race and ethnicity information, there are two drop down boxes that can collect race information. It is unclear from the program logic how the race information is currently being extracted. It is not clear if the race information in the ethnicity dropdown box is being extracted as race information. State needs to explain what is currently being extracted, and to explain changes in the new program logic.

State should remove Hispanic/Latino ethnicity from the possible choices for race. State should consider training workers to ask specifically for race, regardless of ethnicity and to correctly use the "unable to determine" response. Race choices that cannot be mapped to the AFCARS race values, such as "other," must be mapped to blank (missing), not to "unable to determine."

State collects multiple ethnicities.

The frequencies indicate that 25% of children are of Hispanic origin. The State team noted that this percentage is lower than expected.

On the screen that contains the race and ethnicity information, there are two places (a drop down list and a check box) where a worker may select "Hispanic" ethnicity. If a worker selects Hispanic/Latino ethnicity from the drop-down list, the check box for Hispanic/Latino ethnicity is set automatically to "yes." It is unclear from the program code how this information is currently extracted. According to the State, this information should be extracted from the check box. This may be contributing to the under-reporting of Hispanic information because, other ethnicity choices, such as Puerto Rican, Cuban, Dominican, Mexican/Chicano, Mexican/American and Other Spanish/Hispanic do not set the check box to a "yes" response. For accurate AFCARS reporting, all of these ethnicity choices must be mapped to a "yes" response for Hispanic origin.

According to AFCARS frequency report, only 3% of the submitted records have "yes" as a response. Based on the case reviews, it appears that most children have at least one disability recorded in the paper files.

The current program code derives element #10 from element #33. Since element #10 depends on whether "child disability" was selected as a circumstance for removal, it cannot be updated even if a disability is identified later. The data for element #10 should not be based on the circumstances associated with removal. However, it is okay for this to be used as a cross-referencing edit, if the child was removed due to a disability. Then the response to element #10 should be "yes". The State team noted that the new medical profile ("469") is the more appropriate source of professionally diagnosed disability information.

Whether new AFCARS code collects this information from a single LINK source or from more than one data entry screen, the State must ensure that this information can be updated at any time.

If the "469" medical profile becomes the source of this element, then LINK should require caseworkers to complete the "469." The State should consider also modifying LINK to send workers a reminder about updating this information, especially for those cases where disabilities are "Not Yet Determined." Submit documentation on the medical profile ("469") for Children's Bureau review, if used to collect elements #10-15.

Training/screen design should make it clear that this question asks about a professionally diagnosed disability.

Case File Review Findings: 29 of the 53 cases analyzed did not have matching information between AFCARS and the paper file. All should have been reported as "yes" for having a disability.

The AFCARS frequencies for these elements appear to be under-reported. Based on the case review analysis, about half of the sampled case files included documentation on a "emotional disturbed" disability not reported to AFCARS.

Caseworkers currently enter diagnosed disabilities on the LINK screen that also collects circumstances associated with removal. Once removal information has been entered, workers cannot enter more information about diagnosed disabilities, even when this information becomes available at a later time.

State should review Children's Bureau list of medical and psychological conditions mapped to AFCARS disability definitions, to expand the current list of conditions as needed.

This information is collected through a question on a screen for the worker to answer.

State should consider worker training or an on-screen statement to make sure that workers understand that this question relates to both public and private adoptions.

The program code contains a default date of 99991231 (12/31/9999).

State should evaluate the way the information is being extracted for dates of removal, to determine if program code incorrectly extracts other dates instead of the date of first removal from home.

State conducted a conversion to LINK and brought over information on open cases for the years 1993 - 1996.

Case File Review Finding: 24 of the 53 records analyzed did not have matching data, and 2 were not analyzed. Many cases had earlier dates of first removal from home in the paper case files then what was reported in the AFCARS file. Some of these earlier removal dates should have been captured in LINK during the conversion of cases.

There were four records that showed dates of first removal in AFCARS after 1/1/1997. Two of the AFCARS dates were in the year 2000, and the actual removal dates were 1995 and 1996. Two others had AFCARS dates in 1999, and the actual dates were 1994 and 1996.

The amount of difference in the dates reported to AFCARS and the actual dates ranged from 1 day to 2 months (4) and 10 were off by 2 to 12 years.

There were 19 records that had dates of 1st removal after 1/1/1997. Of those, 13 matched, and 5 did not match.

AFCARS frequencies show that 69% of all children have no more than one removal.

See General Requirements Findings regarding "trial home visit." The accuracy of the data for this element is affected by the State practice of reporting children as discharged when they are returned home while still under care, placement or supervision of the State (AFCARS placement setting of trial home visit). If a foster care child still under State supervision returns to foster care from a trial home visit during a court-specified period of time or before six months (if there is no court-specified time period associated with the trial home visit), then the number of removals should not change. (See Children's Bureau's Child Welfare Policy Manual, section 1, AFCARS.)

Case File Review Finding: The number of removals reported to AFCARS is often different (in 28 of 53 cases) than the number of removals based on information in paper files. One record was not included in the analysis.

The information from the paper case files often contained fewer removals then what AFCARS contained. This is because discharges are reported to have occurred in the AFCARS report, when the case file indicates that the child has been continuously in care.

For those with dates of 1^st removal prior to 1/1/97: 14 matched and 21 did not match. Of the 21 that did not match, 2 increased in the number of removals and 18 decreased.

For those with dates of 1^st removal after 1/1/97, all that did not match decreased in the total number of removals.

State must remove default date of 99991231 (12/31/9999).

See notes on the General Requirements Findings form regarding "trial home visit." The accuracy of the data for this element is affected by the State practice of reporting children as discharged when they are returned home while still under care, placement or supervision of the State (AFCARS placement setting of trial home visit). If a foster care child still under State supervision returns to foster care from a trial home visit after either a court-specified period of time or within six months (if there is no court-specified time period associated with the trial home visit), the date of discharge from a previous episode should not change. (See Children's Bureau's Child Welfare Policy Manual, section 1, AFCARS.)

State should assess whether the question on the screen that asks workers "Is this a close of all placements?" is used to identify an incorrect date for this element and for the of date discharge (element #56). State should consider additional worker training to ensure that this information is being entered.

Case File Review Finding: Case review results indicate that the Date of Discharge from Previous Episode reported in AFCARS is often different (in 32 of 53 cases) than the information in paper files. One record was not included in the analysis.

The paper case files confirm that some of the dates reported to AFCARS do not reflect actual discharges from care (probably changes in placements, particularly from paid to non-paid placement settings).

There were 18 records in the pre-1/1/97 group. Of those, 8 matched and 10 did not match.

There were 35 records in the post-1/1/97 group, and 13 matched and 21 did not match. One record showed a date in AFCARS after the end of the report period under review. 17 of the AFCARS records had dates, but the child had never been discharged and remained in the same removal episode.

See General Requirements Findings regarding "trial home visit." The accuracy of the data for this element is affected by the State practice of reporting children as discharged when they are returned home while still under care, placement or supervision of the State (AFCARS placement setting of trial home visit). If a foster care child still under State supervision returns to foster care from a trial home visit during a court-specified period of time or before six months (if there is no court-specified time period associated with the trial home visit), then the date of latest removal should not change. For cases where a specified period of time is not stated in the court orders, or by State policy, and the child re-enters foster care after six months, then this removal begins a new removal episode. (See Children's Bureau's Child Welfare Policy Manual, section 1, AFCARS.)

State should assess the way the information is recorded and extracted for AFCARS reporting. Based on the case reviews, it appears that program code may incorrectly extract and report dates other than the date of latest removal from home. In discussions with the State team, it was proposed that some incorrect dates may reflect placement changes from unpaid to paid placements or vice versa. Because LINK was designed with a focus on paid placements, when a child changes from an unpaid to a paid placement, it may be incorrectly recorded and reported as a new entry into foster care.

The State should also evaluate the impact of the question on the placement screen that asks if this placement is a removal from home, which may also contribute to the suspected over-counting of removals from home.

Program logic contains a default date of 99991231 (12/31/9999).

Case File Review Findings: 27 of the 53 records analyzed did not match, and 5 were not included in the analysis. The paper case files typically contain an earlier date of latest removal than the one in AFCARS report. There were 17 records pre-1/1/97. Of those, 4 matched and 12 did not. In this group, the dates in AFCARS and the actual dates were off by four to seven years. In the post-1/1/97 group, 17 records matched and 15 did not.

The State's AFCARS extraction code creates this date by adding two days to the date of the child's removal date. The State systems staff indicated that LINK does contain a system-generated date stamp, which is not used by the AFCARS mapping and extraction code. The State must change the AFCARS mapping and extraction code to report the system-generated date stamp created when the worker enters the latest removal date (element #21) for the current removal episode into the automated system. The State must ensure that this system-generated date stamp cannot be deleted, changed or replaced after it is created.

Program code contains a default date of 99991231 (12/31/9999).

Information on non-paid placements is not being entered consistently by workers, resulting in gaps in placement information. Also, State practice includes the entry of "retroactive placements," when placement information is entered into LINK up to several months after the placement initially occurred. "Retroactive placements" are likely to result in inaccurate placement information in both LINK and AFCARS reports.

See notes on the General Requirements Findings form.

State does not enter placement change date when a child returns home for "trial home visit". While this is correct according to ACYF-PIQ-94-01 and ACYF-PI-95-01, the policy will be changing. State should make change to record the date a child enters a trial home visit placement setting.

Program code contains a default date of 99991231 (12/31/9999).

Case File Review Finding: 18 of the 53 cases analyzed did not match what was reported in AFCARS. One record was not included in the analysis.

System is designed around payments and, therefore, placement counts may not include all non-paid placements.

"Retroactive placements" may affect the count of placement settings.

The State is counting placements in a way that is not consistent with the AFCARS guidance.

Program code contains a default date of 99991231 (12/31/9999).

Case File Review Finding: The case review results indicated that in 27 of 53 cases, the paper files contain information that would generate a different number of placement settings. Typically, the paper files contained information on many more placements then were reported in AFCARS. This is probably because the current episode is of a longer duration than reported.

The State code "other" is mapped to AFCARS "court ordered". If this code is used for "96-hour holds," then the choice on the screen should be "96-hour hold" and it should be mapped to AFCARS "not yet determined." Once a court order or a voluntary agreement is obtained, the worker should update the information. If the child leaves out-of-home care by the end of the 96-hour hold, then the child is discharged as of that date and the manner of removal for this episode remains "Not Yet Determined."

Also, if a child originally enters care under a voluntary order and then a court order is obtained for care and placement responsibility, then the response to this element should remain "Voluntary."

Case File Findings: 44 of the 53 records did not match what was in the paper file, one was not included in the analysis. There were only eight records that did not match, but all had been voluntary placements and had been reported to AFCARS as court ordered.

If no condition among elements #26-40 has been identified, the State AFCARS extraction code sets element #28 Neglect to "Applies." Otherwise, "Does Not Apply" is the default value for conditions associated with removal.

State should confirm that reasons for removal are collected for children who are removed and placed in settings not paid by DCF, as well as for those children entering DCF-paid placements.

Low frequency count indicates the need for worker training to enter all relevant conditions. State must set the default to blank (missing).

Case File Review Finding: There were several records that did not have all the information in AFCARS that was found in the paper files for the conditions associated with removal. The areas most frequently identified as not being in the AFCARS files were "parent substance abuse", "child's behavior problem", and "caretaker inability to cope."

"Runaway" and "trial home visit" are not included in the extract. The AFCARS frequencies indicate no children in trial home visits (n=0) and one in runaway status, which may reflect a miscode. If the State agency maintains custody on a child who is returned home, then the State must continue to report the child to AFCARS and the placement setting would be "trial home visit." The State must train workers, and modify LINK if needed, to make sure that element #41 accurately reflects the numbers of children on runaway status and on trial home visits.

Workers are not consistently entering placement information in a timely way. Worker entry of placement data only into a narrative has resulted in placement history gaps and incomplete AFCARS reports. The State should train workers to enter placement information into the LINK data fields, as well as in WORD-based narrative fields.

SAFE Homes placements must be included in the AFCARS data extraction. See the General Findings form for Population Requirements.

Before program code is finalized, a list of codes as mapped to AFCARS should be reviewed by the Children's Bureau.

This information is currently not collected in LINK. The current AFCARS extraction program code defaults missing information to "Case plan goal not yet established." The AFCARS frequencies confirm that 100% of the children in care are reported as "Case plan goal not yet established."

Post-site visit findings: The State re-designed this functional area to include the service plan process and collect information on administrative reviews not collected in LINK as of July 2001. The rollout for this change began August 24, 2001, to be phased in by region. The implementation has been pushed back and it is not expected to be implemented statewide until February 1, 2002.

The State code "other" is mapped to "unable to determine."

It appears that if gender information is unavailable, then the State defaults to "Single Female." The mapping forms do not match the program code.

The mapping form includes family constellation. State will decide level of detail it wants to include for this element.

Program code extracts the information from the court disposition screen.

State does record a deceased date for the TPR, if applicable.

Missing information is left blank.

Case File Review Findings: Case review results indicated that not all TPR dates or date of parent's death contained in the paper files have been reported in AFCARS.

Program code extracts the information from the court disposition screen.

State does record a deceased date for the TPR, if applicable.

Missing information is left blank.

Case File Review Findings: Case review results indicated that not all TPR dates or date of parent's decease contained in the paper files have been reported in AFCARS.

The State code "other" is mapped to "single female".

State program code has more detail information regarding family structure, such as "blended family."

Currently if the family structure is identified as "two parent" this is mapped to "married couple" even though this couple may be unmarried.

Missing data defaults to 1960. AFCARS frequencies confirm a large number of 1st Foster Caretakers reportedly born in 1960.

AFCAR frequencies include both extremely young and old foster parents, suggesting the need for additional data entry edit checks to ensure more accurate data entry.

Missing data is defaulted to 1960. AFCARS frequencies confirm a large number of 2nd Foster Caretakers reportedly born in 1960. State must leave missing information blank.

Frequencies show extreme ages, very young and very old foster parents, suggesting the need for additional data entry edit checks to ensure more accurate data entry.

LINK data entry screens and program code allows the collection of multiple races.

The frequencies for race indicate a large percent (18%) of 1st foster caretakers with a race of "unable to determine." The "unable to determine" response is appropriate only if the foster parent refuses to specify a race. The State team agreed that it may also be possible that workers are entering "unable to determine" as the race when the individual is identified as being of Hispanic or Latino origin.

Also, the AFCARS Foster Care Design Document states that if the "cd_race" field on the person table is blank, then substitute the value to indicate "white". The State needs to check the program logic is not mapping missing data to "white". The Design Document is not the current version, as noted by the use of "1" as the AFCARS value for "white". The document "Additions to Detailed Design" does not include this reference.

On the screen that contains the race and ethnicity information, there are two drop down boxes that can collect race information. It is unclear from the program logic how the race information is currently being extracted. It is not clear if the race information in the ethnicity dropdown box is being extracted as race information. State needs to explain how this is currently being extracted, and to explain changes in the new program logic.

State should remove Hispanic/Latino ethnicity from the possible choices for race. State should consider training workers to ask specifically for race, regardless of ethnicity and to correctly use the "unable to determine" response. Race choices that cannot be mapped to the AFCARS race values, such as "other," must be mapped to blank (missing), not to "unable to determine."

State collects multiple ethnicities.

On the screen that contains the race and ethnicity information, there are two places (a drop down list and a check box) where a worker may select "Hispanic" ethnicity. If a worker selects Hispanic/Latino ethnicity from the drop-down list, the check box for Hispanic/Latino ethnicity is set automatically to "yes." It is unclear from the program logic how this information is currently extracted. According to the State, this information should be extracted from the check box. This may be contributing to the under-reporting of Hispanic information because, other ethnicity choices, such as Puerto Rican, Cuban, Dominican, Mexican/Chicano, Mexican/American and Other Spanish/Hispanic do not set the check box to a "yes" response. For accurate AFCARS reporting, all of these ethnicity choices must be mapped to a "yes" response for Hispanic origin.

LINK data entry screens and program code allow the collection of multiple races through a drop-down check box where workers may identify more than one race.

The frequencies for race indicate a large percent (14%) of 2nd Foster Caretakers with a race of "unable to determine." The "unable to determine" response is appropriate only if the foster parent refuses to specify race. The State team agreed that it may also be possible that workers are entering "unable to determine" for race when the individual is identified as being of Hispanic or Latino origin.

State should remove Hispanic/Latino ethnicity from the possible choices for race. State should consider training workers to ask specifically for race, regardless of ethnicity and to correctly use the "unable to determine" response. Race choices that cannot be mapped to the AFCARS race values, such as "other," must be mapped to blank (missing), not to "unable to determine."

On the screen that collects race and ethnicity, LINK provides two ways (a drop-down list and a check box) where a worker may select Hispanic/Latino ethnicity. The drop-down list allows workers to identify more than one ethnicity from many choices. If a worker selects Hispanic/Latino ethnicity from the drop-down list, the check box for Hispanic/Latino ethnicity (also on that data entry screen) is set automatically to "Yes." Other ethnicity choices, however, such as Puerto Rican, Cuban, Dominican, Mexican/Chicano, Mexican/American and Other Spanish/Hispanic do not trigger the same automatic "Yes" response. For accurate AFCARS reporting, all of these ethnicity choices must be mapped to a "Yes" response for Hispanic origin.

State must modify the AFCARS code where appropriate to report the Hispanic origin of individuals identified as Puerto Rican, Cuban, Dominican, Mexican/Chicano, Mexican/American and Other Spanish/Hispanic.

According to the AFCARS frequencies, this AFCARS report includes discharge dates from the four years prior to the current report period.

It was not clear from the AFCARS extraction program documentation whether element #56 is consistently extracted from a legal/court screen. Other observed problems with elements # 20 and 21 (Date of Discharge from Previous Episode and Date of Latest Removal) may also affect element #56. While the case review results showed that there were only a few of the cases (5 of 53) in which the Date of Discharge was different in the paper files from the AFCARS report, not many of these cases had discharges during the report period under review.

Pre-adopt placement settings may be incorrectly recorded and reported as discharges.

Program code contains a default date of 99991231 (12/31/9999).

The State's AFCARS extraction code creates this date by adding two days to the date of the child's discharge date. The State systems staff indicated that LINK does contain a system-generated date stamp, which is not used by the AFCARS mapping and extraction code.

The program code contains a default date of 99991231 (12/31/9999).

According to the AFCARS frequencies, less than half of all cases with a discharge dates also reported a discharge reason (1,288 cases with discharge reasons, compared to 2,702 cases with discharge dates). Such a large discrepancy suggests potential problems with both data element mapping and data entry.

State staff observed that guardianship appears to be under-reported.

According to the AFCARS frequencies, almost half (47%) of all foster care cases receive Title IV-E assistance.

State should verify whether the information being submitted is based on "eligibility" versus "reimbursed" and confirm that this information is being reported accurately for all cases.

Since 47% of the cases are reported to receive Title IV-E foster care assistance (element #59), there should be a similar percent of children eligible for Title XIX. However, the AFCARS frequencies show that no foster care child is currently eligible for Title XIX.

State must confirm the AFCARS code is initialized to blank (missing). State should also consider whether worker training is needed.

Case File Review Finding: 18 of the 53 records analyzed did not match, and the reviewer found that this information did apply.

This element is derived based on the responses to elements #59, 61, and 62. However, the State should assess whether this captures the information being asked-that is, does the case receive other types of assistance in addition to those listed in elements #59-64.

State should change the mapping so that Railroad Compensation is mapped to element #65.

The State team observed that it was unlikely that nearly 20% of foster children require a monthly payment of more than $2000. Also according to the frequencies, 66% of all foster care cases required no monthly foster care payment.

The program logic appears to include retro payments for retroactive placements. The retro payment is generated when a placement is retroactively entered into the system, enabling a single multi-month payment to be made. Retro payments may account for the finding that 10 percent of element #66 monthly amounts are over $5,000. State should consider eliminating the practice of retroactive placements and payments.

The State expects element # 66 to be generated based on a batch program that determines the monthly payment amount based on the age group and provider type. State must evaluate the program code and make changes as needed.

See Notes in General Requirements finding.

State will change the way Record Numbers are reported to AFCARS with the 2002A file. Record Numbers will be reported in the current way through 2001B.

The AFCARS extraction code default is set to "No." If there is an adoption subsidy the value is changed to "yes".

This element is derived from elements #31, 34 and 35. If #31, 34, &35 are blank, then element #4 is set to "No." If #31, 34, and 35 are set to "Applies," then element #4 is set to "Yes."

See notes for element #35.

State team noted that there is a monthly program that checks every case for a birth date. Every month the program runs and shows everybody that had a birth date and checks service type by age group. It will create a new placement for them so that worker does not need to do it.

By the time a child reaches subsidy, it is unlikely that the date of birth will be wrong.

Even with a default set for missing Date of Birth, the State is confident that default information is being updated with the actual date of birth when the child is placed, long before adoption occurs.

Program code currently defaults to 1980/01 for missing date of birth.

The State team is confident that child's gender is updated, if needed.

State code maps an "Other" to "unable to determine" which is not a valid code for element #6.

Default is set to "Male."

In the adoption cases, none of the demographic information is carried over from the foster care case management screens. This has the potential for errors when information has to be re-entered. The State should consider modifying the system to automatically transfer foster care demographic information over to the adoption case record.

"Other" is mapped to "White."

The frequencies for race indicate a large percent (40%) of adopted children with a race of "unable to determine." The "unable to determine" response is appropriate only if a child was abandoned or if the child, parent and/or caretaker refused to specify race. If the State does not modify the system to automatically transfer foster care demographic information over to the adoption case record, State should consider training workers to use the "unable to determine" response correctly.

See also foster care element #8.

The frequencies indicate that 17% of adopted children are of Hispanic origin and that no adopted child (n=0) has an "unable to determine" Hispanic origin.

AFCARS extraction code currently defaults to "No." The State must initialize to blank, if this information is to be re-entered. As recommended above in the element #7 comments, State should consider modifying the system to automatically transfer foster care demographic information to the adoption case record.

See foster care element #9.

Comparing the frequency count of this element (100%= "No") to element #35 (n=36 cases receiving monthly subsidy), there is an inconsistency between the reporting of these elements. State team acknowledged that there is a problem with the reporting of subsidy and special needs information.

The "certification of special needs" screen has the capability to collect the basis of special needs (element #10) specifying a primary, secondary, etc., basis. This information can be used to populate the response to element #9. The existing program code is going to this screen for the information. However, workers are not completing this information.

Default is set to "No."

See comments in adoption element #9.

The frequency count indicates that 100% of adoption cases are "Not Applicable." The default is set to "Not Applicable."

State should also evaluate the list of special needs on the drop-down box for those choices such as "mental disability" which need to be more clearly stated to ensure that the choice can be mapped with certainty to the appropriate AFCARS value.

For elements #11-15, the State should consider modifying the system to carry this information over from the foster care screens. Once the change has been made, it will allow workers to update this information as needed.

"Other" is mapped to "Does not apply."

It appears that the program logic is extracting the current marital status. There is a question on the "certification for special needs" screen, but workers may not be completing this screen.

In order to obtain a TPR, the agency is required to know the marital status of the mother at the time the child was born.

There is not an option for the worker to select "unable to determine."

Default is set to "No."

Program code initializes this element to blank, but then missing TPR dates are calculated by subtracting 9 months from the date on the adoption decree. The new program code should extract actual TPR dates and not derive the missing dates based on the date of the adoption decree.

TPR dates are entered on the "legal screen."

Program code initializes this element to blank, but then missing TPR dates are calculated by subtracting 9 months from the date on the adoption decree. The new program code should extract actual TPR dates and not derive the missing dates based on the date of the adoption decree.

TPR dates are entered on the "legal screen."

Program code initializes this element to blank.

The State central office staff enters this information upon the receipt of the adoption decree. It is not being populated from a discharge of episode screen. The State does this as an additional quality assurance step. It is not clear though if the "reason for discharge" of "adoption" is getting entered at the time of the finalization of the adoption.

"Other" is mapped to "Unable to determine" (which is not a valid AFCARS code).

The program code looks at foster parent constellation, which contains more detail then just "married, unmarried, or single." State team noted that they expect to drop the family constellation information gathering approach and use the family structure information. (See also comments for Foster Care elements #44 and 49 on caretaker and foster caretaker family structures).

See Foster Care element #8 for more information about how race data is currently collected in LINK and Children's Bureau's recommended changes.

Default of race set to "White" may be evidenced in the frequencies (54%="White," 36%="Black," and 8%="Unable to determine."

See Foster Care element #9 for more information about how Hispanic origin data is currently collected in LINK and Children's Bureau's recommended changes.

The AFCARS frequency count reports 18% of cases with Adoptive Mother's Hispanic Origin = "Yes."

This information is in the adoption decree screen/table. The State is able to report all relationships that apply.

State has the options of "other" and "unknown" on the picklist. If these choices are selected they are mapped to "does not apply". This would mean that there would be no information on the relationship between the child and the adoptive parents.

State has only two to three cases a year that are non-public agency adoptions.

Mapping is based on the type of provider that is entered by workers using a drop-down list. The State plans to re-evaluate the provider list.

"Other" is mapped to "public agency."

State team believes that the information is not being correctly extracted from the system.

If a child receives a Medicaid only subsidy, the response to this element should be "yes".